#12 Gemini 3 Flash Preview
low- Cost
- $0.007
- Time
- 12.1s
- Tokens
- 2,289 tok
Google
Release: 2025-12-17
Tested on: 2026-05-22 00:26
google/gemini-3-flash-preview::low
~500B total (~20B active)MoEClosedEstimated
Summary
Gemini 3 Flash Preview scores 8.6 on AI BENCHY and ranks #12. It has 10.0 reliability, a 81.7% pass rate, $0.106 total cost, and 5.86s average response time.
What makes Gemini 3 Flash Preview unique: Its total benchmark cost is unusually low for its score range. It is notably fast compared with similar models.
Model facts
Researched on 2026-08-12
- Parameters
- ~500B total (~20B active)
- Architecture
- MoE
- Availability
- Closed
- License
- -
Best estimate from public evidence; the vendor did not disclose every value. Sparse MoE is documented for the family; counts are estimated.
8.6
Consistency
9.6
10.0
$0.106
Total Output Tokens
29,491
Total Input Tokens
0
Input Price
$0.500 / 1M
Output Price
$3.000 / 1M
Flaky tests
1
Flaky tests had mixed outcomes across runs (at least one pass and one fail).
Hamster playing table tennis
Prompt: Create a detailed SVG illustration of a hamster playing table tennis.
Run history
| Tested on | Score | Reliability | Tests Correct | Total Cost | Compare |
|---|---|---|---|---|---|
| 2026-08-14 02:42 Re-test | 7.4 | 10.0 | $0.185 | Compare | |
| 2026-07-16 22:51 New test added | 7.4 | 10.0 | $0.177 | Compare | |
| 2026-06-04 13:59 New test added | 8.4 | 10.0 | $0.111 | Compare | |
| 2026-05-22 00:26 Suite changed | 8.6 | 10.0 | $0.106 | Current run | |
| 2026-04-11 01:44 First recorded run | 8.8 | N/A | $0.091 | Compare |
This run used a different benchmark suite. Keep suite changes in mind when reading historical movement.
Run comparison
| Run | Benchmark coverage | Score | Consistency | Reliability | Tests Correct | Flaky tests | Total Output Tokens | Total Input Tokens | Total Cost | Response Time (avg) |
|---|---|---|---|---|---|---|---|---|---|---|
| 2026-05-22 00:26 · Suite changed | 60/60 attempts | 8.6 | 9.6 | 10.0 | 16/20 | 1 | 29,491 | 0 | $0.106 | 5.86s |
| 2026-08-14 02:42 · Re-test | 66/66 attempts | 7.4 | 9.2 | 10.0 | 16/22 | 2 | 40,776 | 123,687 | $0.185 | 6.51s |
| Difference | — | +1.3 | +0.4 | 0.0 | 0 | -1 | -11285 | -123687 | -$0.079 | -643ms |
Benchmark coverage differs: 60/60 attempts (Target: 3 repeats per test) versus 66/66 attempts (Target: 3 repeats per test). Totals and repeat-sensitive metrics are not directly comparable.
These two runs used different benchmark suites, so the deltas reflect both model changes and suite changes.
Charts
Choose the first model, then click a second model to open a side-by-side page.
Top Models by Score
Score vs Total Cost
Response Time (avg)
Score vs Response Time (avg)
Total Output Tokens
Score vs Total Output Tokens
Quick Compare
Gemini 3 Flash PreviewlowvsGrok 4.6lowGemini 3 Flash PreviewlowvsClaude Sonnet 4.6noneGemini 3 Flash PreviewlowvsQwen3.8 2.4T A95BhighGemini 3 Flash PreviewlowvsGLM 5.3highGemini 3 Flash PreviewlowvsLongCat 2.0mediumGemini 3 Flash PreviewlowvsKAT-Coder-Pro V2.5lowGemini 3 Flash PreviewlowvsGPT-5.6 LunamediumGemini 3 Flash PreviewlowvsGemini 3.1 Flash LitemediumGemini 3 Flash PreviewlowvsKimi K2.7 CodemediumGemini 3 Flash PreviewlowvsQwen3.7 Plusnone
Category Breakdown
| Category | Score | Consistency | Tests Correct |
|---|---|---|---|
| Anti-AI Tricks | 10.0 | 10.0 | |
| Coding | 7.3 | 10.0 | |
| Combined | 3.0 | 10.0 | |
| Data parsing and extraction | 10.0 | 10.0 | |
| Domain specific | 5.3 | 7.2 | |
| General Intelligence | 10.0 | 10.0 | |
| Instructions following | 9.9 | 10.0 | |
| Puzzle Solving | 10.0 | 10.0 | |
| Tool Calling | 10.0 | 10.0 | |
| Trivia | 10.0 | 10.0 |